a new technique to estimate net groundwater use across large … · 2015-09-22 · groundwater has...

10.1007/s10040-004-0394-5

Hydrogeology Journal Official Journal of the International Association of Hydrogeologists

© Springer-Verlag 2004

10.1007/s10040-004-0394-5

Paper

A new technique to estimate net groundwater use across large irrigated areas by combining remote sensing and water balance approaches, Rechna Doab, PakistanMobin-ud-Din Ahmad1 , Wim G. M. Bastiaanssen2 and Reinder A. Feddes3

(1) Hydrologist and Remote Sensing Specialist, International Water Management Institute (IWMI), PO Box 2075, Colombo, Sri Lanka

(2) Water Watch, General Foulkesweg 28, 6703 BS Wageningen, The Netherlands

(3) Agrohydrology and Groundwater Management, Department of Water Resources, Wageningen University, Nieuwe Kanaal 11, 6709 PA Wageningen, The Netherlands

Mobin-ud-Din Ahmad Email: [email protected] Phone: +94-11-2787404 Fax: +94-11-2786854

Received: 29 August 2003 Accepted: 15 September 2004 Published online: 6 November 2004

Abstract Over-exploitation of groundwater resources threatens the future of irrigated agriculture, especially in the arid and semi-arid regions of the world. In order to reverse this trend, and to ensure future food security, the achievement of sustainable groundwater use is ranking high on the agenda of water policy makers. Spatio-temporally distributed information on net groundwater use—i.e. the difference between tubewell withdrawals for irrigation and net recharge—is often unknown at the river basin scale. Conventionally, groundwater use is estimated from tubewell inventories or phreatic surface fluctuations. There are shortcomings related to the application of these approaches. An alternative methodology for computing the various water balance components of the unsaturated zone by using geo-information techniques is provided in this paper. With this approach, groundwater recharge will not be quantified explicitly, but is part of net groundwater use, and the spatial variation can be quantitatively described. Records of routine climatic data, canal discharges at major offtakes, phreatic surface depth fluctuations, and

http://www.springerlink.com/media/99a0hcwgqn3rypqhlk1p/contributions/l/9/y/2/l9y25e2g0ghbm45j_html/fulltext.html (1 of 25)8-11-2005 9:37:32

mailto:[email protected]

10.1007/s10040-004-0394-5

simplified information on soil textural properties are required as input data into this new Geographic Information System and Remote Sensing tool. The Rechna Doab region (approximately 2.97 million ha), located in the Indus basin irrigation system of Pakistan, has been used as a case study. On an annual basis, an areal average net groundwater use of 82 mm year–1 was estimated. The current result deviates 65% from the specific yield method. The deviation from estimates using tubewell withdrawal related data is even higher.

Keywords Remote sensing - GIS - Water balance - Groundwater management - Net groundwater use - Recharge - Irrigation management - Sustainability - Rechna Doab, Pakistan

Resumen La sobre-explotación de recursos de agua subterránea amenaza el futuro de la agricultura de riego, especialmente en las regiones áridas y semi-áridas del mundo. Para revertir esta tendencia, y para garantizar seguridad alimentaria futura, la meta del uso sostenible del agua subterránea se encuentra alto en la agenda de los políticos. Información espacial y temporal en cuanto al uso neto de agua subterránea- i.e. la diferencia entre las extracciones de agua de pozos entubados para riego y recarga neta- se desconoce frecuentemente a la escala de cuenca hidrográfica. Generalmente, el uso de agua subterránea se estima a partir de inventarios de pozos o fluctuaciones de superficies freáticas. Existen deficiencias en relación con las aplicaciones de estos enfoques. En este artículo se aporta una metodología alternativa para calcular los diferentes componentes del balance hídrico de la zona no saturada utilizando técnicas geoinformativas. Aunque con este enfoque no se cuantifica de manera explícita la recarga de agua subterránea, la cual es parte del uso neto de agua subterránea, puede describirse cuantitativamente la variación espacial. Para esta nueva herramienta de Sistemas de Información Geográfica y Sensores Remotos se requieren datos de entrada como registros rutinarios de datos climáticos, descargas de canales en salidas principales, fluctuaciones de profundidades de superficies freáticas, e información simplificada de las propiedades texturales de los suelos. Se ha utilizado como estudio de caso la región Rechna Doab (aproximadamente 2.97 millones ha), localizada en el sistema de riego de la cuenca Indus de Pakistán. Se ha estimado un uso promedio areal anual de agua subterránea de 82 mm año–1. El resultado obtenido difiere en un 65% del método de productividad específica. La diferencia en relación a estimados provenientes de extracciones en pozos entubados es aún mucho más alta.

Résumé La surexploitation des ressources en eau souterraine menace le futur de l agriculture irrigué, spécialement dans les zones arides et semi-arides du monde. De manière à renverser la tendance, et d assurer la sécurité alimentaire, l utilisation durable des eaux souterraines est devenue une priorité dans l agenda des politiques de l eau. La distribution spatio-temporelle de l usage net de leau souterraine (la différence entre l ‹eau pompée et la recharge nette) est rarement connue à l échelle d un bassin versant.

Conventionnellement, l utilisation des eaux souterraines est estimée à partir des données de rabattement ou les données de fluctuation du niveau de la nappe phréatique. Il y a des défauts dans ces approches. Une méthodologie alternative pour calculer les différents composants de la balance hydrologique est présentée dans cet article. Avec cette approche, la recharge des eaux souterraines ne sera pas quantifiée de manière explicite, mais sera considérée comme une part de l utilisation nette en eau


10.1007/s10040-004-0394-5

souterraine, et la variation spatiale peut être décrite quantitativement. Les chroniques des données climatiques, les débits du réseau hydrographique majeur, les fluctuations de la surface de la nappe phréatique, et des données basiques sur la texture du sol sont nécessaires et sont rentrées dans un nouveau Système d Information Géographique et outil de télédétection. La région de Rechna Doab au Pakistan, environ 2.97 millions d hectare, localisée dans le bassin irrigué de l Indus, a été utilisé comme cas d étude. Sur base annuelle, l utilisation nette de l eau souterraine est estimée à 82 mm. en moyenne. Le résultat obtenu diffère de 65% du résultat de la méthode du débit spécifique. La différence avec le résultat obtenu en observant le rabattement des puits est encore plus élevée.

IntroductionMany large nations rely on irrigated land for more than half of their domestic food production. On irrigated farms, two or three crops per year can be grown and yields are usually high; therefore, the spread of irrigation has been the key to the previous century s rise in food production. The other side of the story is that two thirds of the world s diverted fresh water is being used for irrigation—with an appreciable contribution coming from groundwater resources. Most of the 750–800 billion m3 year–1 of global groundwater withdrawals are used for agriculture (Shah et al. 2000). During the last 10 to 20 years, there has been a significant increase in the utilization of groundwater resources for agricultural irrigation because of their widespread distribution and low development costs (Clarke et al. 1996). Groundwater has been the heart of the green revolution in agriculture across many Asian nations, and has permitted cultivation of high value crops. Today, the United States, China, India and Pakistan are the biggest consumers of groundwater and its use is still increasing (Postel 1999). In the Punjab of Pakistan, use of groundwater for irrigation has rapidly increased, which was mainly initiated in the 1960s with the launch of Salinity Control and Reclamation Projects (SCARPs). Under this program, a number of large-capacity public pumps (0.084–0.14 m3 sec–1) were installed to control waterlogging and provide supplementary irrigation supplies. Thereafter the Government of Pakistan encouraged the installation of private pumps for irrigation. Presently, the greatest proportion of groundwater supplies comes from thousands of small-capacity (0.028 m3 sec–1 or less) private tubewells, which was minimal in the initial stages of the development of this resource. Reliability of timing and supply, and better control over volumes of water applied are just few of the many reasons why farmers have adopted groundwater irrigation on a massive scale (Scott and Shah 2004). The rapid development of tubewells is a clear indication of the current level of farmer reliance on groundwater to irrigation.

In most irrigation systems, the sustainability of groundwater is impacted by two factors: (1) rise of the phreatic surface into the root zone and (2) declining phreatic surface in aquifers that are overused. The major reason for a rising phreatic surface is seepage from irrigation canals and irrigated fields. In fresh groundwater areas (total dissolved salts less than 500 mg L–1), depletion of the aquifer and fall of the phreatic surface are caused by unplanned over-utilization of groundwater. If extraction of groundwater exceeds replenishment, aquifer levels will drop with adverse consequences to entire rural economies and the livelihood of farmers. This drop in groundwater levels can lead to deterioration of groundwater quality due to saltwater intrusion from saline zones. Persistent reliance on such groundwater in irrigated areas has also resulted in the transport of salts from deep aquifers into the root zone resulting in secondary salinity and sodicity. To avoid these undesirable scenarios, it is necessary to assimilate groundwater pumpage and recharge data in solid aquifer exploitation plans.


10.1007/s10040-004-0394-5

Adequate management of groundwater systems is feasible only when all groundwater flow terms are known. This can be achieved by means of numerical groundwater flow models that simulate flows on the basis of locally measured geo-hydrological properties and hydraulic heads. The set-up of a sophisticated groundwater flow model and its calibration is a rather time-consuming effort and is often overlooked because of the absence of sufficient resources and data. Groundwater flow models are, therefore, not commonly used in the South Asian irrigation context.

Groundwater extractions are computed using simpler techniques. For large irrigated areas, the statistical method (Utilization Factor U F-Method)

is used to define the relationship between tubewell extraction and tubewell utilization time, as well as site characteristics (NESPAK-SGI 1991; Maupin 1999). Tubewell utilization time is usually estimated from electricity/fuel usage bills or through field surveys. These field survey data are sometimes biased because certain countries have water quotas or volumetric water pricing. The water table fluctuation or specific yield method is based on the changes of water storage in the saturated zone, with the basic assumption that all outflow is related to withdrawals. The shortcomings of these conventional approaches are:

– The utilization factor method reflects pumping only, and recharge is disregarded. The many small and large tubewells are operated by different rules, and an average extraction is difficult to establish.

– The specific yield method requires a dense network of piezometers and accurate measurements of phreatic levels that are often absent—besides it ignores lateral groundwater flow, which makes it difficult to get a reliable picture of spatial variations at the regional scale.

Water policymakers are now aware of the value of groundwater, and are committed to undertake proper action and even pursue lawsuits to ensure sustainable exploitation. The error margins related to the field data are, however, unacceptably high for implementing a sound groundwater policy. This paper attempts to develop a methodology for regional-scale applications, which relies on the combined use of remotely sensed information and GIS techniques for the estimation of net groundwater use for agriculture—without the inclusion of complex groundwater models. This method can be considered as a tradeoff between sophisticated groundwater flow modelling and the application of simplified analytical solutions. Some qualitative comparisons with the conventional methods are made for the sake of understanding the ball park figures of the differences.

Materials and methodsDefinitions

For groundwater management, quantitative information on water exchange between the unsaturated and the saturated zone is essential, i.e.

, and tubewell irrigation I tw (See Fig. 1). The percolation rate reaching the saturated zone is defined as recharge.


10.1007/s10040-004-0394-5

Throughout this paper, the flux is referred to as capillary rise, also if this water does not reach the root zone. Then net recharge rate q nr

to the saturated zone is often defined as:

Fig. 1 Schematisation of different soil water fluxes in a phreatic aquifer (Icw is canal water irrigation, Pn is net precipitation, ET a is actual evapotranspiration, I

tw is tubewell irrigation, is recharge, is capillary rise, Q in and Q out lateral in-flow and lateral out-flow in the saturated zone respectively

(1)

which describes the net amount of water vertically conveyed to the phreatic surface through the unsaturated zone.


10.1007/s10040-004-0394-5

In this paper, the difference between tubewell extractions I tw and q nr is defined as net groundwater use I ngw, being mathematically expressed as:

(2)

Description of the study area

This investigation was conducted in Rechna Doab (Fig. 2), which is located in the heart of the Indus River basin irrigation system in Pakistan.

Rechna Doab is the interfluvial area between the Chenab and Ravi Rivers. It lies between longitude 71°48 E to 75°20 E and latitude 30°31 N to

32°51 N. The gross area of this Doab is approximately 2.97 million ha, with a maximum length of 403 km and a maximum width of 113 km, including 2.3 million ha of cultivated land. It is one of the oldest and most intensively developed irrigated areas of the Punjab, Pakistan. The area falls in the rice-wheat and sugarcane agro-ecological zones of the Punjab province. Rice, cotton and forage crops dominate the summer season (kharif). Wheat and forage are the major crops in the winter season (rabi). In some parts, sugarcane is also cultivated, which is an annual crop.


10.1007/s10040-004-0394-5

Fig. 2 Canal irrigation network in Rechna Doab, Pakistan. Upper Chenab Canal (UCC), Bambanwala-Ravi-Bedian-Depalpur (BRBD), Marala-Ravi (MR) Link, Qadirabad-Balluki (QB) Link, Lower Chenab Canal (LCC), Trimu-Sadhnai (TS) Link and Haveli Canal

The climate is characterized by large seasonal fluctuations of air temperatures and rainfall. The summer is long and hot, lasting from April through September, with maximum air temperature ranging from 21 to 49 °C. Winter lasts from December through February, with maximum air temperature ranging from 25 to 27 °C and sometimes falling below zero at night. The annual precipitation is about 400 mm. The monsoon elapses from June to September and accounts for about 75% of annual rainfall. The climate records of Faisalabad along with the general crop calendar are shown in Fig. 3.


10.1007/s10040-004-0394-5

Fig. 3 Cropping calendar, monthly precipitation rate and reference crop evapotranspiration rate ET o computed according to Penman-Monteith for Faisalabad

The Rechna Doab soils consist of alluvial deposits transported by the Indus River and its tributaries. The soils textures are predominantly medium to moderately coarse, with favourable permeability characteristics and show a similarity throughout the area.

Due to scanty and erratic rainfall, successful agriculture is only possible in Rechna Doab through irrigation. The canal irrigation system (Fig. 2)


10.1007/s10040-004-0394-5

was introduced in 1892 with the construction of the Lower Chenab Canal (LCC). Presently, almost 2/3 of Rechna Doab is fed by a perennial canal system; i.e. the irrigation water flows constantly into a secondary (distributary) and tertiary (watercourse) canal system as long as there is need for water and sufficient flow in the rivers. A normal flow period per year is about 340 days. Non-perennial irrigation systems receive water during the kharif, which are restricted to the upper Chenab Canal (UCC) and Marala-Ravi Link (MR Link) canal. The outlets (moghas) from the distributaries are not gated and are designed to deliver a fixed quantity of water when the canals are flowing at full capacity.

The design flows in the distributaries are based on the historical size of the command area. This system was designed to spread a limited amount of canal water over the entire area supporting a cropping intensity of approximately 65%. The original design objective of the irrigation development was to protect against crop failure and to prevent famine. However, in the last two to three decades cropping intensities have increased up to 150%, being enabled by additional supplies from groundwater extraction. On a weekly or 10-day rotation period (locally called

warabandi ), each farmer is allotted a fixed quantity of canal water proportional to his land holdings. To ameliorate the impact of crop water deficits, public and private tubewells were installed on a large scale. Canal and ground water are now commonly used in conjunction.

Traditional methods for the estimation of groundwater use

Due to unplanned and rapid development of private tubewells, no mechanism is in place to calculate electricity or fuel usage of tubewells in Pakistan. Therefore NESPAK-SGI (1991) has worked out utilization factors U F (ratio of daily tubewell working hours to number of hours in a

day) through field surveys in different areas of Pakistan. U F can be used to compute I tw from the number of tubewells N tw in an irrigated area A

and their average discharge capacity Q tw as:

(3)

This statistical approach is known in Pakistan as the Utilization Factor U F-method.

The Saturated zone water balance method is based on the transient water balance of the saturated zone:


10.1007/s10040-004-0394-5

(4)

where Q in and Q out represents lateral groundwater inflow and outflow respectively, S y specific yield, dh change in phreatic surface and dt

change in time. Net groundwater use I ngw calculation through Eq. (4) requires data on the phreatic surface fluctuation dh, specific yield S y and

lateral groundwater flows Q in and Q out. In groundwater modelling studies lateral flows are determined from piezometer measurements of the

height of the phreatic surface above a certain reference level, knowing the transmissivity of the aquifer (Boonstra and Bhutta 1995). Phreatic surface measurements are not generally available at small time intervals. Therefore conventionally, at larger time steps, e.g. a year or more, Q in

and Q out may be assumed to be equal and then the rise or fall of the phreatic surface is only attributed to , i.e. net groundwater use:

(5)

This method is known as the specific yield S y-method. The accuracy and validity of the results is based on the reliability and density of the

observation points (often less than desirable for a dense network with a fine mesh) as well as the assumption that lateral groundwater flows can be ignored.

Geo-information techniques for the estimation net groundwater use

The use of geo-information techniques is a new method to calculate net groundwater use for agriculture. Whereas the S y-method studies the

saturated zone, this approach describes the water balance and the changes of the unsaturated zone. Net groundwater use I ngw must be determined

as a residual term from water balance analysis of the unsaturated zone over a specific period dt (Fig. 1), which reads as:


10.1007/s10040-004-0394-5

(6)

where

(7)

where dWu represent the change in the unsaturated zone storage and in the volumetric soil moisture content. Note that z1 and z2 differ which

implies that Eq. (7) is valid for changing water table conditions. Combining Eqs. (1), (2) and (6), yields a new expression to compute net groundwater use I ngw as a function of land surface processes that can be obtained nowadays from remote sensing and GIS techniques:

(8)

Equation (8) excludes the need to explicitly solve lateral in- and out-flows. ET a can be calculated using satellite imagery and routine

meteorological data using a remote sensing parameterization of the surface heat fluxes. Several models are available (Kustas et al. 2003 for a recent review). The surface energy balance algorithm for land (SEBAL) developed by Bastiaanssen et al. (1998, 2002, personal communication) has been used in this case study. Since there is a direct link between actual evapotranspiration ET a, potential evapotranspiration ET p and root

zone volumetric soil moisture content at various stages in the growing season (see non-ideal circumstances in Allen et al. 1998), soil

moisture can be estimated on a pixel-to-pixel basis from ET a and ET p. Using the evaporative fraction of the energy balance, Scott et al. (2003)


10.1007/s10040-004-0394-5

made two-weekly estimates of soil moisture in the root zone for the total Indus Basin system. Together with bi-annual measurements of phreatic surface, that describes the moisture conditions in the capillary fringe, unsaturated zone soil moisture storage W u is computed. Ahmad and

Bastiaanssen (2003) developed a new method for this purpose.

On an annual basis in Rechna Doab, net precipitation rate P n is not a major water balance component. Therefore, monthly estimates of

precipitation on a pixel-wide basis were obtained from linear interpolation of gauge readings collected at 8 meteorological stations throughout Rechna Doab. Since Rechna Doab is relatively flat and irrigation fields have bunds, runoff has been ignored. Canal water diversions to all major irrigation canal systems of Rechna Doab have been measured at the head works by the Punjab Irrigation department and Water And Power Development Authority (WAPDA). With the lack of high-resolution satellite images covering the entire Rechna Doab for a complete annual cycle (October 1993- October1994), canal water distribution was disaggregatated using the spatial surfaces of National Oceanic Atmospheric Administration-Advanced Very High Resolution Radiometer based Normalized Difference Vegetation Index (NDVI) maps. NDVI was thus used as a determinant for canal water availability, which has been done before by Menenti et al. (1989). Equation (8) can neither quantify recharge

or net recharge q nr, but the net draft I ngw is the most important for management purposes. This whole approach is more elaborated by

Ahmad (2002).

Results and DiscussionNet groundwater use estimates from geo-information techniques

Spatially distributed values

The maps showing the spatial distribution of the various water balance components derived from remote sensing and geo-information techniques are presented in Fig. 4. Spatial distributions of ET a, varying from 600 to 1,100 mm year–1, are found in the cultivated areas across Rechna Doab.

A large tract with higher ETa exists in the rice growing area in the upper Rechna Doab. These spatial variations in actual evapotranspiration rate

ET a can be attributed to different cropping systems, the conditions of the atmospheric boundary layer and the amount of irrigation water supplied,

including groundwater. Figure 3 has demonstrated that reference evapotranspiration can be as high as 240 mm month–1. If the irrigation canal network does not provide sufficient surface water resources, in fresh groundwater quality areas, farmers will shift to groundwater irrigation. The peak evapotranspiration occurs in the months of June to August from rice areas due to high ambient temperatures and monsoon rains. The lowest evapotranspiration occur in the month of January that is related to low ambient temperatures and to canal closing in Punjab for routine maintenance.


10.1007/s10040-004-0394-5

Fig. 4 Maps showing the distribution of annual irrigation rate with canal water I cw, net precipitation rate P n, actual evapotranspiration rate ET a and change in the unsaturated zone soil moisture storage dWu from October 1993 to October 1994 in Rechna Doab


10.1007/s10040-004-0394-5

Spatial variability in annual I cw between canal commands is mainly the result of the non-perennial system (little or no water supplies in rabi) in

the upper Rechna Doab. Despite the low canal water supplies, the upper Rechna Doab (served by Marala-Ravi MR Link, Bambanwala-Ravi-Bedian-Depalpur BRBD Link and Upper Chenab Canal UCC) is the most intensively cultivated area. Crop water requirement is met by an additional supply from groundwater extraction by tubewells, with some contribution from precipitation. Upper Rechna Doab, in particular the MR canal command, receives 2 to 3 times more precipitation than the middle and lower part of the Doab. The highest rate of irrigation with canal water I cw was estimated in the Haveli Canal, because there are large uncertainties in the gross irrigated areas as reported by the irrigation

department and WAPDA (calculated from canal command coverage). The value of I cw varies between 150 and 1,100 mm. The variation in I cw,

P n and ET a causes positive as well as negative changes in soil moisture storage Wu in Rechna Doab.

Finally, for each pixel, net groundwater use I ngw was estimated by means of Eq. (8). First monthly I ngw values were calculated and then

accumulated to annual values (Fig. 5). Positive I ngw values in Fig. 5 represent net groundwater use, negative values represent net replenishment.

The data shows that annually an amount of 300 to 600 mm is net extracted, thus 0.8 to 1.6 mm day–1. The highest I ngw is observed in the UCC

and BRBD, which are areas containing non-perennial canals. Other fragmented pockets of high I ngw are in the LCC (East), which have higher

annual ET a because of rice cultivation or higher cropping intensities. Scattered patterns of groundwater replenishment (negative values of I ngw)

are also observed across the LCC system, particularly in the head reach with a rice–wheat cropping system. But, most of the groundwater replenishment occurs in the command areas of MR and Haveli. The replenishment in MR is mainly because of a higher net precipitation rate P n,

while in Haveli it is because of higher canal water supplies. Most of the replenishment occurs during the monsoon months when canals are flowing at peak discharge. This water is not necessarily lost as groundwater, as it may be used elsewhere at a later time.


10.1007/s10040-004-0394-5

Fig. 5 Distributed annual net groundwater use I ngw, Eq. (8), from October 1993 to October 1994, as computed using remote sensing and geo-information techniques in Rechna Doab

Accuracy assessment

The estimated values of net groundwater use I ngw from geo-information techniques are prone to errors due to the uncertainty in the assessment of

each input water balance component. Thus it is essential to compute the influence of errors in each individual component, ET a, P n, I cw and dWu/

dt, on overall computation of I ngwfrom Eq. (8) and see whether the errors in individual components will cancel or propagate.

To understand the error behaviour under different agro-climatic and hydrological conditions in Rechna Doab, two pixels representing cotton–wheat and rice–wheat cropping rotation, are selected for detailed investigation. These pixels represent deep and shallow phreatic surface conditions respectively. The annual estimated values for different water balance components from geo-information techniques are presented in


10.1007/s10040-004-0394-5

Table 1. ET a is one of the largest water balance components in Rechna Doab and obviously errors in its computation will have a major influence

on I ngw. Recent studies in Rechna Doab indicate that the accuracy of time-integrated SEBAL ET a varies from 0 to 10% at field scale, and 5% at

the regional scale (Bastiaanssen et al. 2002). Similarly for the same area the accuracy in dWu/dt was found to be around 6% (Ahmad and

Bastiaanssen 2003). The expected accuracy of I cw and P n is about 15% but this could not be validated at the 1 km grid size because of the

unavailability of data at appropriate scale in Rechna Doab. Due to the absence of a time series field for these selected locations, random data series are generated for ET a, P n, I cw and dWu/dt with ±10, ±15, ±15 and ±6% deviations from its estimated values, respectively.

Table 1 Annual water balance components estimated from geo-information techniques at selected 1 km grid in the cotton–wheat system at Faisalabad (73°249.8 E 31°23 26.2 N) and rice–wheat system at Pindi Bhattian (73°20 50.2 E 31°52 34.2 N), in Rechna Doab

Water Balance Component (mm year–1) Faisalabad cotton–wheat system Pindi Bhattian rice-wheat system

ETa 870 810

Pn 195 370

Icw 410 565

dWu/dt –45 85

Using these generated data series I ngw is computed using Eq. (8) for 1,000 different combinations of random input parameters. For both locations,

similar deviations are observed. Finally, the absolute deviation I ngw is estimated from geo-information techniques using random data series and is

plotted against its probability of occurrence (Fig. 6). The figure shows that the maximum error that could occur in the geo-information techniques

is 190 mm year–1 for similar hydrological conditions. However, there is a 70% chance that the error I ngw will be within the range of 75 mm year–

1. The average error (50% probability) is less than 50 mm year–1. The areas where the rate of net groundwater use is high—300 mm year–1 or more—i.e. the priority areas for groundwater management, this error is within the acceptable range (25% or less).


10.1007/s10040-004-0394-5

Fig. 6 Distribution of absolute deviation of net groundwater use I ngw from geo-information techniques estimates using random input data series in Rechna Doab

Aggregated values at canal command level

Monthly estimates of I ngw in different canal commands of Rechna Doab, as calculated from the geo-information based water balance analysis,

are presented in Fig. 7. Positive and negative values of I ngw were observed in all canal commands during both the rabi and kharif. The highest

values are found in February and May (>150 mm month–1) due to the high crop water requirement (more pumping, positive I ngw). The peak in

February can be explained by the flowering stage of wheat and in May for the land preparation for cotton (see cropping calendar of Fig. 3). The


10.1007/s10040-004-0394-5

values of I ngw in MR, UCC, and BRBD are much higher than in other canal commands during the rabi, since little or no canal water is being

diverted to these canal commands. During the monsoon months (July, August and September), I ngw was often negative, which implies that

groundwater systems are replenished in most of the canal commands. Accumulated results on an annual basis reveal that the UCC, BRBD and LCC use considerable amounts of net groundwater: 170, 108 and 85 mm year–1 respectively. In MR and Haveli, the aquifer was replenished with 405 and 201 mm year–1 respectively, which is a significant quantity. Although not specifically investigated, it is plausible that excessive groundwater flows from MR to BRBD and UCC; the same for Haveli to LCC.

Fig. 7 Monthly net groundwater use Ingwin different canal commands of Rechna Doab. Net groundwater use occurs during crop sowing periods while replenishment occurs during monsoon periods. Marala-Ravi Link MR, UCC Upper Chenab Canal, BRBD Bambanwala-Ravi-Bedian-Depalpur Canal, LCC Lower Chenab Canal and Haveli Canal

Records of routine climatic data, canal discharges at major offtakes, fluctuations in phreatic level, and simplified information on soil textural properties are required as input data into this new geographic information system and remote sensing tool. The results are less sensitive to phreatic


10.1007/s10040-004-0394-5

surface depth than the reliance of the specific yield method on this data. Hence, under the collection efforts practiced by the Pakistani line agencies, there is sufficient field data available to execute the new geo-information technology approach.

Net groundwater use estimates from conventional techniques

Utilization factor UFmethod

Tahir and Habib (2000) used U F, N tw and Q tw data (GOP 1994) for the approximation of I tw, Eq. (3), in the Punjab. As tubewell statistics were

available for large aggregated districts, they first calculated I tw for the district and then transformed it to a canal command scale, according to the

proportion of the district to the specific command area. Estimated annual I tw values for different canal commands of Rechna Doab are

summarized in Table 2. The highest I tw was found in the non-perennial canal commands that are located in the upper Rechna Doab: MR, UCC

and BRBD (see Fig. 2). This area is in the rice–wheat agro-climatic zone of the Punjab and most groundwater withdrawals was used for rice cultivation.

Table 2 Annual estimates of groundwater withdrawal for irrigation I tw (Eq. (<equationcite>3</equationcite>) for the year 1993–94, using utilization factors U F

in different canal commands of Rechna Doab. Marala-Ravi Link MR Link, UCC Upper Chenab Canal, BRBD Bambanwala-Ravi-Bedian-Depalpur Canal, LCC Lower Chenab Canal and Haveli Canal (partly after Tahir and Habib 2000)

Canal command

MR UCC BRBD LCC Haveli

Itw (mm year–1) 932 629 729 460 540

Specific yield Sy method

Across Rechna Doab, the depth of the phreatic surface was measured twice a year, pre- and post-monsoon, with a nodal network of 981 piezometers by the SCARP Monitoring Organization (SMO) of Water and Power Development Authority (WAPDA) of Pakistan. They also conducted pumping tests at 47 different locations to calculate the specific yield S y for Rechna Doab (Khan 1978). These point data of specific

yield and piezometric levels for October 1993 and October 1994 were acquired from WAPDA and interpolated using the kriging method.


10.1007/s10040-004-0394-5

Thereafter, the annual change in phreatic surface depth dh was calculated from the maps representing the depth of phreatic surface in October 1993 and October 1994. After combining the dh and S y maps, the annual net groundwater use I ngw was computed using Eq. (5) from October

1993 to October 1994 (Fig. 8), thus ignoring lateral inflow. The I ngw phreatic surface have positive and negative values due to rising and falling.

The spatial patterns of I ngw follow the kriged phreatic surface level variations, and are thus somewhat artificial. Moreover, lateral transfer is

entirely ignored, being a shortcoming of this approach.

Fig. 8 Annual net groundwater use I ngw (Eq.(5) (obtained from measurement of the phreatic surface on October 1993 and October 1994 and specific yield S y

data in Rechna Doab). Positive values show net groundwater use and negative values show net replenishment


10.1007/s10040-004-0394-5

Comparison between conventional and geo-information techniques

A comparison between Fig. 5 and Fig. 8 shows that the geo-information based analysis of Fig. 5 produce I ngw values that give different spatial

shapes, with more abrupt changes than the gradual patterns of Fig. 8. The effect of change in land use, soil types and irrigation practices on I ngw

is more pronounced in the geo-information based analysis than in an analysis that is based on kriged S y and phreatic surface measurements. The

geo-information based water balance analysis has a wide spread of values as compared to the S y method due to the full description of agricultural

and irrigation practices. In the positive range, the values from the geo-information techniques are significantly higher than the values from the S y

method, and thus the values of the S y-method are systematically underestimated.

The main difference between the two methods is that the conventional approach ( S y-method) neglects groundwater lateral flow. The S y-method,

which ignores lateral groundwater flow, yielded a 65% underestimation of net groundwater use. Net lateral flows to Rechna Doab from October 1993 to October 1994 amounted to 1193 million m3 year–1 (which is equivalent to a water layer of 53 mm year–1). The largest inflows were found in the LCC and UCC respectively. This also confirms that in highly permeable aquifers, such as in Rechna Doab, lateral groundwater flow cannot be ignored. Any localized change in the phreatic surface depth caused by groundwater withdrawals or irrigation activity tends to be offset by exchanging groundwater flows.

It is interesting to compare values for groundwater withdrawal by tubewells I tw (a traditional way of referring to groundwater use for irrigation in

Pakistan) using the U F-method, Eq. (3), with net groundwater use I ngw obtained by both the geo-information and the S y-Method (Table 3).

Except for Haveli, the geo-information and S y-method show similar trends of annual net groundwater use in all commands. The net groundwater

use from the UF method is significantly higher (550%) than the results obtained from the geo-information techniques. This can be partially

explained by the net recharge component that is not considered in the U F-Method. Moreover the use of one average utilization factor for different

types of private tubewell is, with different discharge capacities, may be one of the other reasons that resulted in this overestimation.

Table 3 Comparison of annual net groundwater use I ngw computations from the geo-information techniques (Eq. (8) and the S y-method (Eq. (5) with groundwater withdrawal by tubewells I tw calculated using the utilization factors U F (Eq. (3) in the different canal commands of Rechna Doab from October 1993 to October 1994. Q in and Q out represent respectively the groundwater lateral in-flow and out-flow. (Marala-Ravi MR Link, UCC Upper Chenab Canal, BRBD Bambanwala-Ravi-Bedian-Depalpur Canal, LCC Lower Chenab Canal and Haveli Canal.). Negative values express and positive values are net replenishment groundwater use


10.1007/s10040-004-0394-5

Canal commands Geo-information Sy-Method UF-Method Geo-information Sy-Method UF-Method Geo-information

Ingw

mm year–1

Ingw

mm year–1

Itw

mm year–1

Ingw

106 m3 year–1

Ingw

106 m3 year–1

Itw

106 m3 year–1

Qin- Qout

106 m3 year–1

MR –405 –54 +932 –264 –35 +606 –229

UCC +170 +24 +629 +732 +103 +2702 +629

BRBD +108 +13 +729 +191 +23 +1292 +168

LCC +85 +34 +460 +1289 +511 +6979 +778

Haveli –201 +65 +540 –115 +37 +308 –152

Rechna Doab (106 m3 year)–1) +1833 +639 +11882 +1193

Rechna Doab (mm year–1) +82 +28 +529 +53

Summary and ConclusionsIn this paper, net groundwater use I ngw was computed using remote sensing and geo-information system methods for calculating the unsaturated

zone water balance, Eq. (8). The annual average water balance components in Rechna Doab are summarized in Fig. 9. In Rechna Doab an amount of 82 mm is skimmed off from the groundwater resources every year.


10.1007/s10040-004-0394-5

Fig. 9 Annual water balance from October 1993 to October 1994 over the Rechna Doab as estimated from geo-information techniques, where I ngw is net groundwater use, I cw is rate of irrigation with canal water, P n is rate of net precipitation, ET a is rate of actual evapotranspiration, I tw is rate of irrigation with groundwater withdrawal from tubewells, q nr is net recharge, and Q in and Qoutare lateral in-flow and out-flow in the saturated zone (dWu/dt=+31 mm)

The specific yield method produced a 65% lower net groundwater use as compared to the geo-information techniques. This difference is explained by groundwater lateral flows that bring in a net amount of 53 mm, which prevents the phreatic surface from declining too rapidly. This implies that in highly permeable phreatic aquifers the groundwater withdrawal by tubewells or net recharge from irrigated fields cannot be inferred from occasional monitoring of the phreatic surface. Despite limited phreatic surface measurements are useful for trend analysis, they cannot be used for a quantitative assessment of groundwater withdrawals for agriculture. This finding has significant consequences for groundwater use, exploitation and policy making in general.

Also, groundwater withdrawal computations using utilization factors, Eq. (3), differ greatly from the net groundwater use calculations, Eq. (8). In this method, return flows towards the aquifer are not considered, which is an important process in groundwater irrigated agricultural systems.

Even at canal command level, geo-information techniques provide more accurate results than the conventional techniques. This is because geo-information based net groundwater use combined with the S y-method depicts whether the rise or fall of the phreatic surface can be ascribed to net

lateral groundwater flow or to vertical interactions. Such information is of strategic importance when considering localized or regional options for sustainable groundwater management.

Hence, geo-information techniques provide a comprehensive, efficient, and standardized opportunity of quantifying net groundwater use for agriculture with a minimal need of field data. The technique is especially suitable for large scale applications such as for river basins with rapidly falling phreatic surfaces.

Acknowledgements The authors wish to express their gratitude to the Punjab Irrigation Department (PID), Pakistan Meteorological Department (PMD), and Water And Power Development Authority (WAPDA) of Pakistan for making available the required information.

References


10.1007/s10040-004-0394-5

Ahmad MD (2002) Estimation of net groundwater use in irrigated river basins using geo-information techniques: a case study in Rechna Doab, Pakistan. PhD Thesis, Wageningen University, Wageningen, 144 pp (http://www.gcw.nl/dissertations/3334/dis3334.pdf)

Ahmad MD, Bastiaanssen WGM (2003) Retrieving soil moisture storage in the unsaturated zone from satellite imagery and bi-annual phreatic surface fluctuations. Irrig Drainage Sys 17(3):141–161

Allen RG, Pereira LS, Raes D, Smith M (1998). Crop evapotranspiration, guidelines for computing crop water requirements, FAO Irrigation and Drainage Paper 56, Food and Agriculture Organization of the United Nations (FAO), Rome, 300 pp Bastiaanssen WGM, Ahmad MD, Chemin Y (2002) Satellite surveillance of evaporative depletion across the Indus Basin. Water Resour Res 38(12)1273 Bastiaanssen WGM, Menenti M, Feddes RA, Holtslag, AAM (1998) A remote sensing surface energy balance algorithm for land (SEBAL), part 1: formulation. J Hydrol 212–213:198–212 Bastiaanssen WGM, Noordman EJM, Pelgrum H, Davids G, Thoreson BP, Allen RG (2005) Use of SEBAL model with remotely sensed data to improve water resources management under actual field conditions. ASCE Irrigation and Drainage Engineering (in press) Boonstra J, Bhutta MN (1995) Groundwater recharge in irrigated agriculture: the theory and practice of inverse modelling. J Hydrol 174:357–374

Clarke R, Lawrence AR, Foster SSD (1996) Groundwater: a threatened resource. UNEP environment Library Series 15, UNEP, Geneva, http://www.unep.net/

GOP (1994) Agriculture machinery census. Agriculture Census Organization, Statistics division, Government of Pakistan (GOP), Lahore Khan MA (1978) Hydrological data, Rechna Doab: lithology, mechanical analysis and water quality data of testholes/testwells, vol I. Publication no. 25, Project Planning Organization (NZ), Pakistan Water and Power Development Authority (WAPDA), Lahore Kustas WP, Diak GR, Moran MS (2003) Evapotranspiration, Remote Sensing of. Encyclopedia of Water Science, Dekker, New York, pp 267–274 Maupin MA (1999) Methods to determine pumped irrigation-water withdrawals from the Snake River between upper Salmon Falls and Swan Falls dams, Idaho, using electrical power data, 1990–95. U.S. Geological Survey Water-Resources Investigation Report 99–4175, USGS, Littleton, Colorado, 20 pp Menenti M, Visser TNM, Morabito JA, Drovandi A (1989) Appraisal of irrigation performance with satellite data and georeferenced information. In (eds) Rydzewski and Ward, Irrigation, theory and practice. Proc. Int. Conf., 12–15 September 1989, Univ. of Southampton, Southhampton, pp 785–801


http://www.gcw.nl/dissertations/3334/dis3334.pdf

http://dx.doi.org/10.1023/A:1025101217521

http://dx.doi.org/10.1016/0022-1694(95)02724-6

http://www.unep.net/

10.1007/s10040-004-0394-5

NESPAK-SGI (National Engineering Services Pakistan Ltd.-Specialist Group, Inc.) (1991) Contribution of private tubewells in the development of water potential. Final report, Planning and Development Division, Ministry of Planning and Development, Islamabad Postel S (1999) The pillars of sand: Can the irrigation miracle last? The Worldwatch Institute, Washington, DC Scott CA, Bastiaanssen WGM, Ahmad MD (2003) Mapping root zone soil moisture using remotely optical imagery. ASCE Irrig Drainage Eng 129(5):326–335

Scott CA., Shah T (2004) Groundwater overdraft reduction through agricultural energy policy: insights from India and Mexico. Water Resour Devel 20(2):149–164

Shah T, Molden D, Sakthivadivel R, Seckler D (2000) The global groundwater situation: overview of opportunities and challenges. International Water Management Institute, Colombo, Sri Lanka, 21 pp. (http://www.iwmi.org)

Tahir Z, Habib Z (2000) Land and water productivity: Trends across Punjab canal commands. Working Paper no. 14, Pakistan country series no.3, International Water Management Institute (IWMI), Colombo, Sri Lanka, 35 pp, (http://www.iwmi.org)


http://dx.doi.org/10.1061/(ASCE)0733-9437(2003)129:5(326)

http://dx.doi.org/10.1080/0790062042000206156

http://www.iwmi.org/

http://www.iwmi.org/

a new technique to estimate net groundwater use across large … · 2015-09-22 · groundwater has...

Documents